Letter pubs.acs.org/macroletters
Ammonium Bicarbonate Transport in Anion Exchange Membranes for Salinity Gradient Energy Geoffrey M. Geise,†,‡ Michael A. Hickner,*,† and Bruce E. Logan‡ †
Materials Science and Engineering and ‡Department of Civil and Environmental Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, United States S Supporting Information *
ABSTRACT: Many salinity gradient energy technologies such as reverse electrodialysis (RED) rely on highly selective anion transport through polymeric anion exchange membranes. While there is considerable interest in using thermolytic solutions such as ammonium bicarbonate (AmB) in RED processes for closed-loop conversion of heat energy to electricity, little is known about membrane performance in this electrolyte. The resistances of two commercially available cation exchange membranes in AmB were lower than their resistances in NaCl. However, the resistances of commercially available anion exchange membranes (AEMs) were much larger in AmB than in NaCl, which would adversely affect energy recovery. The properties of a series of quaternary ammonium-functionalized poly(phenylene oxide) and Radel-based AEMs were therefore examined to understand the reasons for increased resistance in AmB to overcome this performance penalty due to the lower mobility of bicarbonate, 4.59 × 10−4 cm2/(V s), compared to chloride, 7.90 × 10−4 cm2/(V s) (the dilute aqueous solution mobility ratio of HCO3− to Cl− is 0.58). Most membrane resistances were generally consistent with the dilute solution mobilities of the anions. For a few key samples, however, increased water uptake in AmB solution reduced the ionic resistance of the polymer compared to its resistance in NaCl solution. This increased water uptake was attributed to the greater hydration of the bicarbonate ion compared to the chloride ion. The increased resistance due to the use of bicarbonate as opposed to chloride ions in AEMs can therefore be mitigated by designing polymers that swell more in AmB compared to NaCl solutions, enabling more efficient energy recovery using AmB thermolytic solutions in RED.
R
membranes in AmB versus NaCl electrolytes, we measured the ionic area resistance of two commercially available cation exchange membranes (CEMs) and two commercially available anion exchange membranes (AEMs) using these two salts (Figure 1). The area resistances of the AEMs in AmB were greater than those in NaCl. As a result, switching from NaCl to AmB in a RED device could reduce energy efficiency even though the area resistance of the CEMs in AmB is less than that in NaCl. AEMs that do not have substantially increased ionic resistance with AmB compared to NaCl (i.e., AEMs that behave differently from those commercial AEMs shown in Figure 1) are desirable for use in AmB-based salinity gradient technologies such as RED to overcome the differences in membrane resistance based on the properties of the electrolyte. The differences in resistance measured using the two electrolytes reported in Figure 1 can be rationalized, to a first approximation, as resulting from differences in the dilute solution mobilities of ammonium and bicarbonate ions compared to sodium and chloride ions. The dilute solution mobilities of NH4+ and Na+ are 7.71 × 10−4 cm2/(V s) and 5.18 × 10−4 cm2/(V s), respectively, so for a CEM that transports
ecent interest in salinity gradient energy technologies to address global energy needs, such as reverse electrodialysis1−4 (RED), capacitive energy extraction based on Donnan potential5 (CDP), and capacitive reverse electrodialysis6 (CRED), has encouraged the study of ion-containing polymers to optimize ion exchange membranes for these processes.4,7,8 These technologies rely on membranes that are capable of selectively transporting ions under the influence of an electric field.7,8 The ionic resistance of the membranes in these processes is a critical parameter as large resistances reduce the energy efficiency of the devices.1,2,9,10 Many researchers have reported the ionic resistance of polymers measured using aqueous NaCl1,11 because sodium and chloride are the predominant ions in many natural water sources.12 Thermolytic salts, such as aqueous ammonium bicarbonate (AmB), are being considered for salinity gradient energy production because these solutions permit closed-loop conversion of heat energy to electricity.4,13,14 Little is known, however, about AmB transport in different types of polymer membranes. The ionic resistance of ion-containing polymers measured using NaCl solutions may not be representative of the ionic resistance of those materials in AmB electrolytes. The efficiency of AmB-based technologies, such as RED for waste heat recovery, critically depends on controlling rates of ammonium and bicarbonate transport through ion-containing polymers. To demonstrate that ion transport was different for © XXXX American Chemical Society
Received: July 2, 2013 Accepted: August 16, 2013
814
dx.doi.org/10.1021/mz4003408 | ACS Macro Lett. 2013, 2, 814−817
ACS Macro Letters
Letter
Figure 1. Area resistance at room temperature of commercially available ion exchange membranes (Selemion CMV and AMV and Membranes International, Inc. CMI-7000 and AMI-7001) measured using either 0.5 mol/L NaCl or AmB solutions. Figure 2. Structures of the anion exchange polymers, which are shown in their “as-prepared” counterion form, considered in this study.
predominantly cations the resistance ratio of sodium chloride to ammonium bicarbonate (rm,NaCl/rm,AmB) might be expected to be 1.49, i.e., the mobility ratio of NH4+ to Na+.15 This ratio is favorable for AmB-based RED because the resistance of the CEM is lower in AmB compared to NaCl. The dilute solution mobilities of HCO3− and Cl− are 4.59 × 10−4 and 7.90 × 10−4 cm2/(V s), respectively, so for an AEM that transports predominantly anions the resistance ratio of sodium chloride to ammonium bicarbonate (rm,NaCl/rm,AmB) might be expected to be 0.58, i.e., the mobility ratio of HCO3− to Cl−, which is unfavorable for AmB-based RED because the predicted resistance of the AEM is greater in AmB compared to NaCl.15 The ratios of the resistances of the commercial membranes in NaCl to that in AmB (Figure 1), which can be compared to the values calculated from dilute solution mobility, are 3.7 ± 1.8 (CMV), 1.7 ± 0.1 (CMI-7000), 0.35 ± 0.05 (AMV), and 0.54 ± 0.07 (AMI-7001). The approximate agreement between the ratios of the resistance data presented in Figure 1 and the ratios of the dilute solution mobilities is reasonable given that these commercial membranes are reinforced using a nonconductive material and do not have greatly different water uptake in either electrolyte solution. To understand how to reduce the ionic resistance of AmB in AEMs, we investigated the properties of a range of candidate materials as alternatives to the commercial membranes. A key difference between the commercial samples and our experimental samples was that our membranes were not mechanically restricted by the presence of a reinforcing material. Exposing a nonreinforced polymer to a bicarbonate-containing salt can swell the polymer to a greater extent than would occur in a chloride-containing salt because bicarbonate ions are more highly hydrated than chloride ions.16 Because membranes that absorb more water generally have lower resistance, we wanted to take advantage of this swelling phenomenon to decrease the resistance of the membrane when exposed to the low mobility bicarbonate-containing electrolyte. The polymers selected for the experimental AEMs had aromatic backbones with various levels of quaternary ammonium functionality (Figure 2). Synthesis details for quaternary ammonium functionalized poly(sulfone) (aRadel17) and quaternary ammonium functionalized poly(phenylene oxide) (aPPO18) have been reported previously. Additional details about the polymers, film casting,
and pre-experiment conditioning procedures are provided as Supporting Information. Ionic resistance was measured using a direct current (DC) method19 described in the Supporting Information. For the AEMs shown in Figure 2, we report the thickness-normalized intrinsic resistance (in units of Ω m) because thickness can have a profound effect on a polymer’s area resistance (in units of Ω cm2, used in Figure 1), and as noted, the commercial membranes have a heterogeneous reinforced structure that prevents accurate calculation of intrinsic resistance.9 Data were reported as averages of at least three measurements, and standard error propagation20 was used to estimate experimental uncertainty based on the standard deviations of the ionic resistance and thickness measurements. The intrinsic resistance of several aRadel and aPPO polymers (Figure 3) was greater in AmB (rm,AmB) than NaCl (rm,NaCl),
Figure 3. Intrinsic resistance values for aPPO and aRadel measured in 0.5 mol/L sodium chloride or ammonium bicarbonate. The ratio of the intrinsic resistance measured in sodium chloride to that measured in ammonium bicarbonate is indicated on the plot for each polymer, and the dashed line represents the calculated ratio of the bicarbonate dilute solution mobility to that of chloride. 815
dx.doi.org/10.1021/mz4003408 | ACS Macro Lett. 2013, 2, 814−817
ACS Macro Letters
Letter
when polymers have the same water uptake in each electrolyte their resistances are consistent with the dilute solution mobility of the mobile ion. The resistance ratios of the aPPO-C6D6 and aPPO-C16D4 samples were consistent with their much greater water uptake in AmB than NaCl, resulting in greater membrane swelling. Water uptake data for the commercial membranes (Figure 1) equilibrated in 0.5 mol/L NaCl and AmB are reported and discussed as Supporting Information. Increasing the water uptake of a polymer generally facilitates ion transport,7,8,21,22 and in this case, higher water uptake resulted in lower ionic resistance (i.e., higher conductivity). For aPPO-C6D6 and aPPO-C16D4, greater water uptake in 0.5 mol/L AmB compared to NaCl lowered the resistance of the polymer measured using AmB, thereby resulting in a rm,NaCl/ rm,AmB ratio above the value predicted using dilute solution mobilities (i.e., above the dashed line in Figures 2 and 3). Increased water uptake in a polymer in AmB compared to NaCl could be related to the activity of water in these two electrolytes. Equations of state, such as Flory−Huggins theory,23,24 predict that the volume fraction of water sorbed in (and, thus, the water uptake of) a polymer will be higher when the polymer is exposed to a higher activity of water. As such, one might expect that the water uptake of a polymer would be greater in AmB compared to NaCl based on the slightly higher activity of water of 0.985 in 0.5 mol/L AmB (determined using OLIAnalyzer) compared to 0.984 for 0.5 mol/L NaCl. According to Flory−Huggins theory (see also Supporting Information), this small difference in water activity would produce
